Influence of ACN Content On The Properties of NBR/CNT Nanocomposites
Influence of ACN Content On The Properties of NBR/CNT Nanocomposites
Influence of ACN Content On The Properties of NBR/CNT Nanocomposites
www.kgk-rubberpoint.de
Experimental
Materials
The investigations were carried out using
NBR with different ACN contents (Perbu-
nan® NT 2845, NT 3445 and NT 3945,
Lanxess AG, Germany) and MWCNTs
(NANOCYL NC 7000, Belgium) produced
via Catalytic Carbon Vapour Deposition
(CVD). Commercial CNTs as specified on
NC 7000 product datasheet show an ave-
rage length around 1.5 – 3 µm, carbon
purity 90%, metal oxide 10 % and surface
area 250-300 m2/g. The characteristics of
the NBR are given in Table 1. Other addi-
tives and curing agents were used as re-
ceived.
Sample preparation
Fig. 1: Mixing diagrams for NBRs with 2.5 vol. % CNTs
Master batches of NBR with concentra-
tions of MWCNTs ranging from 0 to 4 vol. %
were prepared by dry melt mixing using 1 Characteristics of the NBRs used
a laboratory internal mixer (Haake Rheo- Product Acrylonitrile content (%) Mooney viscosity ML Density (g/cm3) Tg (°C)
mix 3000) at 50 rpm and variable mixing (1+4) 100 °C
time. Subsequently, a part of mixtures
were milled on two roll mill for 5, 10 and
NBR 28 28 ± 1 45 ± 5 0.98 -27.3
15 minutes. The required quantity of
NBR 34 34 ± 1 45 ± 5 0.99 -20.7
CNTs was weighed in a MBRAN® glove
box and transferred into an internal mi- NBR 39 39 ± 1 45 ± 5 1.00 -13.5
xer using safety masks. Zinc oxide (7 phr)
was added after 10 minutes, stearic acid (BDA 40, Novaconctrol GmbH) in the fre- a first maximum. During the dispersion
(1 phr) was added at 12 minutes and the quency range of 0.1 Hz to 10 MHz. To and distribution stage of the CNTs and
curing system containing Vulkacit Thi- ensure good contact with the electrodes, the torque values decreases asymptoti-
uram (2 phr), Vulkacit NZ/EG (1.5 phr), the samples were sputtered with gold. cally to a stationary value. The addition
Rhenogran S-80 (1 phr) and Vulkalent Dynamic–mechanical properties of of zinc oxide, stearic acid and curatives
E/C (1.5 phr) was introduced into the in- vulcanized samples were measured by were carried out stepwise. For unfilled
ternal mixer after 15 minutes. using the Rheometrics Dynamic Analy- rubbers the mixing energy (determined
The optimum cure time (t90+1 min/ zer-RDA II at room temperature, 1 Hz by integration of the mixing torque over
mm) was determined from the torque frequency in the strain amplitude range time at constant rotor speed increased as
data using a rheometer (Alpha Technolo- from 0.01 to 3 %. a function of ACN content (Fig. 2). The
gies MDR 200 E). The samples were The stress-strain behaviours of the difference can be seen in the plot of the
vulcanized in a hot press at 160 °C and cured composites were investigated mixing energy against the concentration
200 bar for the rheometer optimum. using a universal tensile testing machine of the CNTs. In the range of the experi-
(Zwick/ Roell 2010) at a strain rate of 200 mental errors, the slope of the torque
Characterization mm/min in accordance with DIN 53 504 increase was almost the same for the
The viscosity of the compounds was mea- standards from vulcanized flat sheets. three rubbers used as matrix materials. A
sured by using the Mooney MV 200 E at ‘Shore A’ hardness was measured using corresponding increase in the tempera-
100 °C. The viscoelastic properties of the Zwick Test Durometer A Type BS 06. ture of the mix was observed.
uncured mixtures were measured with a In the range of small strain amplitu-
Rubber Process Analyzer (RPA 2000) at Results and discussion des (0.3 to 10 %) the storage modulus (G’)
80 °C and 1 Hz frequency within the Mixing and the loss modulus (G’’) of the uncured
strain sweep region 0.28-446 %. The nanocomposites were prepared by compounds showed a slightly decreasing
The morphology of the samples was incorporating CNTs into the plasticized plateau. At small strain amplitudes both
examined by using a transmission elect- polymer by dry mixing in the internal G‘ and G‘‘ increased with the ACN con-
ron microscope (TEM, Zeiss Libra®120) on mixer. The evolution of the torque is qui- tent of the rubber and the concentration
ultracut slices of 100 nm at an accelerati- te similar for the different grades of NBR of the CNTs (Fig. 3). It was observed that
on voltage of 120 kV. but varies significantly with the amount the increase in storage modulus G’ was
Electrical properties were recorded on of the CNTs (Fig. 1). Subsequently, the more pronounced than the one of the
cylindrical samples (Ø 20 mm, d=2 mm) addition of CNTs to the plasticized rub- loss modulus G’’ for all types of NBR abo-
using a Dielectric Broadband Analyzer ber the torque increases sharply reaching ve of 1.5 vol. % CNTs, both of which incre-
2 3
Fig. 2: Mixing energy for NBRs as a function of the CNT loading Fig. 3: Mixing energy for NBRs as a function of the CNT loading
ases with the ACN content of 4 CNTs (Fig. 4). The decrease is si-
the rubber. The particular dif- milarly graded for the three ty-
ferences indicate the contribu- pes of NBR. The acceleration of
tion of the specific interac- the cure process by CNTs was
tions of ACN groups with the observed for other rubbers [17]
π-electrons on the surface of Even if the effect might of-
the CNTs but also with polar fer some advantages in redu-
functional groups formed du- cing the optimum cure time,
ring the purification of these there is no reasonable expla-
fillers. At sufficiently large nation for the acceleration of
strain amplitudes, the storage the curing process. For samp-
modulus approaches the lo- les prepared by two steps, in
wer bound of the unfilled rub- internal mixer and then on
ber. In that region, the loss two-roll mill, the torque diffe-
modulus shows a higher value rence decreased. This effect
Fig. 4: Influence of ACN on t90 and incubation time
compared to the storage mo- was attributed to a partial
dulus which can be attributed breakdown of CNTs. However,
to flow processes. When the measure- Vulcanization the cross-linking characteristics (incuba-
ment was repeated after a relaxation An increasing CNTs concentration resul- tion time and t90 time) are not influ-
period of 30 minutes, it was seen that for ted in a change of the curing isotherms enced by a subsequent milling process.
all systems the mechanical modulus re- for the compounds investigated. The ma-
covered almost completely and the path ximum torque difference increased line- Morphology
of the second strain sweep followed the arly with the amount of CNTs. The slope is The TEM examination of the CNTs as de-
original one irrespective of the filler con- somewhat higher in comparison to car- livered revealed agglomerated and partly
tent. This indicated that after deformati- bon black (N330). Both the incubation entangled clusters of tubes of several
on the system relaxed to a similar state time and the t90 – time were slightly re- hundred microns in length (Fig. 5a). From
to that of the post mixing. duced with increasing concentration of closer investigations it was observed
Fig. 5.: a) TEM image of commercial CNTs (50,000x) b) TEM micrograph of CNT/NBR 34 nanocomposite containing 2.5 vol. % CNTs (10,000x)
c) Mechanical breakdown of CNTs (31,500x)
6 7
Fig. 6: Electrical conductivity as a function of frequency for CNT/ Fig. 7: Influence of ACN content on EPT
NBR 34
Fig. 8: a) Influence of ACN and milling on EPT, b) Influence of milling time on EPT, c) Influence of a combined mixing in mixer and two roll mill
on conductivity for NBR 34 with 2.5 Vol. % CNTs
that the NC 7000 CNTs used in this inves- At higher magnification the cut dia- loading increases further, the electrical
tigation have small average diameters of meter of CNTs are clearly seen (Fig. 5c). conductivity increases over several deca-
ca. 9 nm and a relatively narrow diame- Here CNTs exhibited a broad length dis- des. With concentration of 4 vol. % CNTs,
ter size distribution [18]. tribution, due to mechanical breakdown. the conductivity values were reached up
The extent of filler dispersion and the It was observed that high shear rates and to 0.03 S/cm. This level of conductivity
resulting “in rubber structure” morpho- prolonged mixing led to better homo- can not be obtained by using carbon
logy of the CNTs depends on the mixing geneity but also to a mechanical degra- black at considerably higher loadings.
conditions and the polar nature of the dation of the tube length (aspect ratio is Considering the influence of the ACN it
polymer. To visualize the effect of the dry lowered). became obvious that there was a pro-
melt mixing on the morphology of the nounced shift of the electrical percolati-
nanocomposites, ultra-microtomed cry- Electrical properties on threshold (EPT) from 1.8 vol. % to 1.3
osections of cured samples were exami- The electrical conductivity was measu- vol. % for CNTs depending upon an increa-
ned. The nanocomposites reveal a ran- red in the frequency range from 0.1 Hz se of the ACN content from 28 to 39 wt. %
dom distribution of predominantly long up to 10 MHz. as a function of the CNT respectively (Fig. 7).
and principally well-dispersed CNTs (Fig. content (Fig. 6). Starting from a very low Besides a rather small influence of the
5b). There were only a few rather small conductivity level of the unfilled NBR viscosity of the rubbers, this shift can be
entangled agglomerates that could be formed by emulsion polymerization at- mainly attributed to the interaction of
detected in the 2D-images. However, the tributed to residues. A sharp increase of the ACN groups with specific sites on the
average length of the CNTs is considerab- the conductivity level was observed surface of the CNTs. It can be assumed
ly reduced as a result of scission during within the concentration range of 1 to that the double bonds and specially the
mechanical mixing. The large and com- 1.5 vol. % of CNTs. The plateau value in- ACN segments of the rubber provide be-
pact objects shown in Fig 5b are attribu- creased by three orders of magnitudes neficial interaction with π-electron of
ted to ZnO. Therefore, it can be stated but also the slope of electrical conducti- the CNTs, which are in a special electro-
that the dry mixing process is a quite ef- vity versus the frequency of the current nic state due to the pyramidalization of
fective procedure to disperse and distri- decreased, indicating the formation of a C-C bonds. It is likely that the CN-dipoles
bute CNTs in NBR. The 2D projection filler network in which the charge carri- establish additional specific interactions
showed a good picture of the features of ers demonstrate a normal diffusion [16]. with oxygen-containing functional
the filler network. It is worth mentioning that if the CNTs groups on the surface of tubes [21]. Even
9 10
Fig. 9: Influence of CNTs content on G‘ for NBR 34 Fig. 10: Influence of ACN and CNTs content at 0.09 % strain amplitude
12 13
Fig. 12: Young‘s modulus at 5 % strain Fig. 13: Stress strain curves for NBR types
14 15
Fig. 14: Influence of ACN and CNTs content on reinforcing factor Fig. 15: Tensile strength comparison of NBR types
chains to the rigid surface of CNTs could hancement of the stress values up to 200 The slope increases by the content of
slow down polymer dynamics. Thus, a % elongation reflected the transfer of the polar ACN groups of the rubber indica-
higher amount of mechanical energy is stress from the polymer to the interface ting the positive effects of the ACN
dissipated when the material passes the of the embedded CNTs. The reinforce- groups on the level of the mechanical
glass transition temperature. The rein- ment by CNT was clearly seen in the in- reinforcement. In addition it was found
forcement caused by the presence of crease of the stress in the strain region that the reinforcing factor of 4 vol. %
CNTs was clearly seen by an increase in up to 200 % and the increase of the ten- CNTs corresponds to the 14.5 vol. % of CB
G‘ and G‘‘ in the rubbery plateau. sile strength. Compared to the unfilled Concerning the influence of CNTs on ulti-
NBR 34 samples the effect of CNTs resul- mate properties of the compounds it was
Stress-strain behaviour ted in an increase in tensile strength of 7 seen that, the tensile strength increased
The incorporation of CNTs into different MPa and a reduction in strain of 120 % with the CNTs content (Fig. 15). It can be
types of NBR rubbers led to a significant was found by an addition of 4 vol. % CNTs seen that above the EPT or MPT this in-
increase in the Young’s modulus, the (Fig. 13). A similar graded increase was crease is more pronounced and reveal
stress values at medium elongations and found in the other NBR nanocomposites. the contribution of the ACN content.
tensile strength. Considering the concen- By defining a reinforcing factor by the Below percolation threshold limit
tration dependency of the Young’s mo- ratio of the stress value for CNTs filled strength was limited in the range of 1 to
dulus two discrete regimes were obser- system normalized to the stress values of 2 MPa.
ved. Above the percolation threshold the the unfilled system, the impact of the However, the elongation at break de-
slope of the Young’s modulus was signifi- filler can be described quantitatively. On- creases with the filler content for the
cantly higher as a consequence of the ce again, the relationship between the samples with CNTs concentrations below
response of the filler network to small reinforcing factor and the CNTs content the percolation threshold. In this regime
deformations (Fig. 12). Tensile strength can be divided in two discrete regimes. also an influence of the ACN content on
increased with the CNTs content of the Above the MPT the slope of the reinfor- elongation is observed. Above the MPT
rubber nanocomposites and became mo- cing factor as a function of the CNTs vo- the polarity of the rubber plays a minor
re pronounced by the formation of CNTs lume fraction is considerably higher than role on the elongation at break of the
networks in the polymer matrix. The en- below this characteristic value (Fig. 14). CNT/NBR compounds (Fig. 16).
16 17
Fig. 16: Influence of ACN and CNTs content on reinforcing factor Fig. 17: Tensile strength comparison of NBR types
If the accumulated effect of CNTs on NBR by application oriented mixing in [10] X. W. Zhou, Y. F. Zhu, J. Liang, Mater. Res.
the evolution of the stress over the strain the internal mixer and two roll mill. The Bull., 42 (2007) 45+.
is considered as the energy at break, it can properties of the compound are determi- [11] L. D. Perez, L. F. Giraldo, B. L. Lopez, M. Hess,
be shown that this critical material pro- ned by two main factors: the CNT loa- Macro-mol. Symp., 628 (2006) 245.
perty is considerably increased (i) above ding and the ACN content of the rubber. [12] P. - C. Ma, G. Marom et al. Composites: Part
the percolation threshold and (ii) with the Both the electrical and the mechani- A 41 (2010) 1345.
increasing content of nitrile groups. cal percolation threshold occur at very [13] M. Knite, K. Ozols, J. Zavickis, V. Tupureina,
The mechanical properties did not small CNTs loadings (< 1.5 vol. %) and I. Klemenoks; Journal of Nanoscience and
change significantly when the mixes we- depend on the ACN content of the NBR Nanotechnology 9 (2009) 3587.
re prepared in the internal mixer and as well as on mixing conditions. The sa- [14] H. Chougule, H. Zimmermann, R. Schuster,
subsequently processed on the two roll turation conductivity is inversely propor- ICCE-19 (2012), Shanghai , China.
mill for a given time (Fig. 17). It was tional to the percolation threshold. [15] L. Bokobza, Macromolecular Materials and
found that a subsequent processing Significant mechanical reinforcing Engineering, 289 (2004) 607.
stage though milling did not affect the effects are obtained at CNTs loadings [16] A. Das, M. Klüppel, R. Jurk, M. Saphianniko-
values of the tensile strength even above the percolation threshold. va, J. Fritzsche, Lorenz H, et al. Polymer, 49
though the milling time was prolonged. (2008) 52767.
For the system CNT/NBR34 with 1.5 vol. Acknowledgement [17] H. Zimmermann, R. Schuster, KGK 64 (2011)
% CNT, it was found that tensile strength The authors thank to Germany’s federal 40.
values ranges from 7 to 8 MPa for diffe- Ministry of Education and Research [18] H. Chougule, H. Zimmermann, R.H. Schus-
rent mixing conditions. For this small fil- (BMBF) for financial support of the pro- ter, Inno. CNT Jahreskongress (January
ler content no specific effects of the ACN ject “CarbElast” within the “Innovations- 2012), Bayreuth.
content on the ultimate properties were allianz CNT”. [19] R. H. Schuster, H. Chougule, H. Wittek, ACS
observed. There is reason to state that Rubber Division (181st Technical Meeting,
prolonged mixing may result in a break- References April 2012), San Antonio (Tx, USA).
down of the CNTs which could affect the [1] G. Kraus, “Reinforcement of Elastomers” [20] Liu and Z. Cheng, Chem. Rev. 105 (2005)
electrical properties of the filler network Wiley-Interscience, NY (1965). 3643.
and lowers the aspect ratio of CNTs but [2] J. -B. Donnet, R. C. Bansal and M.-J. Wang [21] S. Parro et al. / Physica E 37 (2007) 58.
does not change the volume fraction of “Carbon Black Science and Technology” [22] Hui-Ming Cheng et al. / Carbon 46 (2007)
the polymer-filler interface and hence Marcel Dekker Inc. NY (1993). 2003.
does not directly influence ultimate pro- [3] R. H. Schuster, “Mixing of Rubber Com- [23] Manfred Klüppel et al., Macromol. Mater.
perties considerably. pounds” (A. Limper, Ed.) Ch. 6, 2012. Eng. 294 (2009) 551.
It was found that the “Shore A” hard- [4] L. Bokobza, Polymer, 48 (2007) 4907.
ness of the NBR nanocomposites increa- [5] P. M. Ajayen, L. S. Schadler, P. V. Braun Nano-
ses slightly as a function of ACN and CNT composites-Science and Technlogy, Wiley-
content. Below the percolation threshold, VCH, Weinheim, 2003.
the hardness was not significantly incre- [6] G. Schmid, Nanoparticles, Wiley-VCH, Wein-
ased with ACN content but the conside- heim, 2004.
rable change on the hardness was higher [7] D. Leon, A Manuel, Polymer engg. & Science
at 4 vol. % of CNT content. (2009) 866.
[8] B. Lopez, L. Perez, M. Mesa, Annual Technical
Conclusion Coneferece-ANTEC, Conference proceedings
It has been shown that CNTs can be in- (2005).
corporated and well dispersed in solid [9] J. E. Mark, ACC. Chem. Res., 39 (2006) 881.